1006
Hydrological Sciences–Journal–des Sciences Hydrologiques, 51(6) December 2006
Rainfall and streamflow response to El Niño Southern Oscillation: a case study in a semiarid catchment, Australia
*
SUNIL C. DUTTA1,2 , JOHN W. RITCHIE1, DAVID M. FREEBAIRN1 & G. YAHYA ABAWI2 1 Department of Natural Resources and Mines, PO Box 318, Toowoomba, Queensland 4350, Australia 2 Department of Primary Industries and Fisheries, PO Box 102, Toowoomba, Queensland 4350, Australia
[email protected]
Abstract This paper aims to compare the shift in frequency distribution and skill of seasonal climate forecasting of both streamflow and rainfall in eastern Australia based on the Southern Oscillation Index (SOI) Phase system. Recent advances in seasonal forecasting of climate variables have highlighted opportunities for improving decision making in natural resources management. Forecasting of rainfall probabilities for different regions in Australia is available, but the use of similar forecasts for water resource supply has not been developed. The use of streamflow forecasts may provide better information for decision-making in irrigation supply and flow management for improved ecological outcomes. To examine the relative efficacy of seasonal forecasting of streamflow and rainfall, the shift in probability distributions and the forecast skill were evaluated using the Wilcoxon rank-sum test and the linear error in probability space (LEPS) skill score, respectively, at three river gauging stations in the Border Rivers Catchment of the Murray-Darling Basin in eastern Australia. A comparison of rainfall and streamflow distributions confirms higher statistical significance in the shift of streamflow distribution than that in rainfall distribution. Moreover, streamflow distribution showed greater skill of forecasting with 0–3 month lead time, compared to rainfall distribution. Key words Australia; El Niño Southern Oscillation; Kolgomorov-Smirnov test; Murray-Darling Basin; rainfall forecasting; seasonal forecasts; Southern Oscillation Index; streamflow forecasting; Wilcoxon rank-sum test
Réponse de la pluie et de l’écoulement en cours d’eau à l’Oscillation Australe El Niño: une étude de cas dans un bassin semi-aride d’Australie Résumé Cet article compare le changement de distribution fréquentielle et l’aptitude à la prévision climatique saisonnière de l’écoulement en cours d’eau et de la pluie en Australie orientale à partir du système de Phase de l’Indice d’Oscillation Australe (SOI). De récents progrès dans la prévision saisonnière de variables climatiques ont ouvert des opportunités pour améliorer la prise de décision en matière de gestion des ressources naturelles. La prévision des probabilités de pluie est disponible pour différentes régions d’Australie, mais aucune utilisation de prévisions semblables n’a été développée pour l’approvisionnement des ressources en eau. L’utilisation de prévisions d’écoulements de cours d’eau pourrait fournir une meilleure information pour la prise de décision dans les domaines de l’irrigation et de la gestion des écoulements avec de meilleurs effets écologiques. Afin d’examiner l’efficacité relative de la prévision saisonnière de l’écoulement en cours d’eau et de la pluie, le changement de distributions fréquentielles et l’aptitude à la prévision ont été respectivement évalués avec le test Wilcoxon “rank-sum” et le score de précision de l’erreur linéaire dans l’espace de probabilité (LEPS), pour trois stations de jaugeage du bassin des Border Rivers, au sein du bassin de Murray-Darling en Australie orientale. Une comparaison des distributions de pluie et d’écoulement confirme que la significativité statistique du changement de distribution est meilleure pour l’écoulement que pour la pluie. De plus, l’aptitude à la prévision dans un délai de à 3 mois est meilleure avec la distribution de l’écoulement qu’avec la distribution de la pluie. Mots clefs Australie; Oscillation Australe El Niño; test Kolgomorov-Smirnov; bassin de Murray-Darling; prévision de pluie; prévisions saisonnières; Indice d’Oscillation Australe; prévision d’écoulement en cours d’eau; test Wilcoxon “rank-sum”
* Now at: Infrastructure Planning, Gold Coast Water, PO Box 5042, Gold Coast MC, Queensland 9729, Australia;
[email protected] Open for discussion until 1 June 2007 Copyright © 2006 The State of Queensland (Australia) Dept of Primary Industries and Fisheries
Rainfall and streamflow response to ENSO
1007
INTRODUCTION In Australia, in the last decade, the importance of water management has increased as water management policies have become more complex, recognising the finite nature of the resource. The irrigation industry is important nationally and contributes approximately 25–30% of the gross value of agricultural production (MDBC, 2001). This equates to approximately Au$7.2 billion, or just over 1% of Gross Domestic Product (ABS, 2004). On the downside, the degree of water extraction from river systems has led to environmental degradation, which appear to be worsening as diversions for consumptive uses increase (Murray-Darling Basin Ministerial Council, 1995). Recognition of these problems has led to a change in the direction of water policy and the adoption of a water reform agenda in the 1990s (ARMCANZ, 1995). The fundamental intention of this reform process is to promote sustainability by examining the allocation and management of Australia’s water resources. Despite a high level of water infrastructure investment, Australia’s irrigated agricultural production still remains variable, driven by highly variable weather sequences (Podbury et al., 1998). Australia’s climate is amongst the most variable in the world resulting in high variation of rainfall and streamflow (Finlayson & McMahon, 1991, Chiew & McMahon, 2003). In recent years, research has examined the drivers and impacts of climate variability in Australia. It has been found that the El Niño Southern Oscillation (ENSO) phenomenon influences the climatic pattern over much of the world, including eastern Australia (Stone & Auliciems, 1996; Latif et al., 1994; Nazemosadat & Cordery, 1997; CPTEC, 2006; Chiew, 2006a). Streamflow variability and its relationship with ENSO has also been studied in Australia and elsewhere in the world and shows a useful relationship in many parts of USA and Australia (Kuhnel et al., 1990; Simpson et al., 1993; Dracup & Kahya, 1994; Amarasekera et al., 1998; Wang & Eltahir, 1999; Hamlet & Lettenmaier, 1999; Franks, 2003). Coupled ocean–atmosphere models have been developed and run for a number of years in predicting ENSO several months in advance (BOM, 2005). The Centre for Ocean-Land-Atmosphere studies (COLA) has recently developed an anomaly coupled prediction system, using sophisticated dynamic ocean and atmosphere models that produce skilful forecasts of tropical Pacific sea-surface temperature anomalies (SSTA) up to 18 months in advance (Kirtman & Shukla, 2006). The effect of ENSO on the rainfall patterns in Australia has been described previously (Pittock, 1975; McBride & Nicholls, 1983; Allan, 1988; Lough, 1991; Murphy & Ribbe, 2004). This research has proposed the use of ENSO based streamflow forecasting systems with water managers and the irrigation community seen as prime users (Abawi et al., 2001). The combination of increased water demand and awareness of potential environmental degradation stemming from over-use has led to the search for ways to improve the management of river systems and associated water infrastructure. The ability to provide reliable probabilistic forecasts of water supply has the potential to improve water management through more judicious allocations in forecasted “drier” periods, and advances in allocation in forecast periods of abundance. The ability to forecast both rainfall and streamflow poses a problem for potential users of forecast information in the area of water management. While there are relationships between rainfall and streamflows, it would be useful to determine which Copyright © 2006 The State of Queensland (Australia) Dept of Primary Industries and Fisheries
1008
Sunil C. Dutta et al.
forecast is likely to be most useful in making water management decisions. Although rainfall is the source of soil moisture and streamflows, its occurrence at a particular point is not necessarily representative of the whole of the catchment. Additionally, rainfall events do not always result in runoff to rivers unless demands of evapotranspiration, interception, infiltration, surface storage, and channel losses are satisfied (Schwab et al., 1966). The lack of spatial coverage of rainfall data and the complexity of catchment characteristics are likely to confound any simple relationships between streamflow and point rainfall. On the other hand, streamflow is an aggregation of the rainfall-runoff processes over the sub-catchment. It can be hypothesised that the integrating nature of streamflow lends itself to linkages with the regional ENSO phenomenon, and may show higher forecasting ability than rainfall. In this study the hypothesis that the SOI phase system will be more useful in forecasting streamflow than rainfall, will be explored. Also, since rainfall excess is a small proportion of the water balance (
